Violent behavior is of substantial social concern. Among potential biologic factors, serotonin (5-HT) neurotransmission has been widely implicated in aggressive behavior. A rich literature suggests that serotonergic (5-HT) activity is inversely associated with aggressive behavior in both human and non-human subjects. The human evidence for this relationship, however, consists almost exclusively of non-experimental research. Therefore, a causal relationship between serotonergic status and aggression in humans has not yet been clearly demonstrated. Theorists have proposed that serotonin influences aggressive behavior by altering the threshold at which an organism responds to provocative stimuli. According to this perspective, raising 5-HT levels should decrease aggressive responding to perceived provocation. At this time, however, no human experimental evidence for the interactive effects of provocation and 5-HT activity on aggression exists. The first aim of this study is to demonstrate a causal relationship between serotonin activity and aggressive behavior in humans. This aim will be addressed by determining if aggressive behavior is reduced after experimentally increasing brain levels of 5-HT. Specifically, 132 subjects (66 individuals with a history of impulsive aggression and 66 normal controls) will be randomly assigned to receive either an inactive placebo capsule or 0,5 mg/kg or 1.0 mg/kg d- fenfluramine (d-FEN). D-FEN has been shown to reliably increase indices of 5-HT functioning, decrease hostile ideation in humans, and some experimental evidence exists for the antiaggressive effects of this drug in animals. Approximately four hours after drug administration, aggressive responding will be observed using a laboratory paradign that has substantial empirical evidence supporting its validity. The second aim of the study is to determine if provocation moderates the relationship between serotonin and aggression. Therefore, provocation by an adversary during an aggressive encounter will be manipulated as a within-subjects factor. It is hypothesized that aggressive behavior will be attenuated by d-FEN compared with placebo, especially under conditions of high-provocation. This effect is expected to be greater in individuals with a history of impulsive aggressive behavior compared with normal controls.